Intrathecal administration of the acetylcholine esterase-inhibitor neostigmine produces analgesia in human by increasing the endogenous neurotransmitter acetylcholine at the spinal cord level [1, 2, 3]. Central muscarinic cholinergic receptors are involved in this analgesic effect of spinal acetylcholine . Preclinical studies have shown that acetylcholine receptors exist at the peripheral nerve level and that cholinergic agents induce peripheral analgesia by accumulation of cyclic guanine monophosphate (GMP) at the nociceptor level via generation of nitric oxide [4, 5, 6, 7]. A recent clinical study in patients undergoing arthroscopic meniscus repair showed that intraarticular administration of neostigmine produced a moderate but significant analgesic effect . Furthermore, it has been shown that peripheral nerves could be able to facilitate muscarinic receptor-mediated signal transduction . Based on these data suggesting a peripheral cholinergic antinociceptive effect via peripheral cholinergic receptors, we made a hypothesis that neostigmine administered at a peripheral nerve level would increase the duration of postoperative analgesia. Therefore, in this prospective, randomized, double-blind, placebo-controlled study we assessed the efficacy of peripherally administered neostigmine on duration of analgesia and side-effects when added to lidocaine for axillary brachial plexus block.
After obtaining approval from the Institutional Review Board and informed consent we studied 34 ASA I or II patients undergoing elective ambulatory carpal tunnel release performed by the same surgeon. Exclusion criteria included contraindication to axillary brachial plexus block such as abnormal coagulation or local cutaneous infection, or patients with peripheral neuropathies. All patients received rectal ketoprofen 100mg and oral premedication with hydroxyzine 1mg kg−1, 1h before surgery.
Standard intraoperative monitoring was used. The patient's arm was placed at 90° to the trunk with the elbow flexed at 90°. All the axillary plexus blocks were performed by the same anaesthesiologist (A.V.E.). A 22-gauge insulated needle (Vygon Laboratoire Pharmaceutique, Ecouen, France) was used to locate the median nerve. A peripheral nerve stimulator (Stimuplex Dig; Braun, Melsugen, Germany) was used for precise location of the nerves. After the position of the needle was judged adequate - when a current output <0.7mA still elicited a slight distal motor response-all patients were administered 1.5% lidocaine 450 mg and epinephrine 5 μg mL−1 through the needle.
Patients were allocated to one of two groups using a table of random numbers and were unaware of their treatment group. Either saline (0.9% NaCI) 1 mL (group S, n=16), or neostigmine 500 μg (Roche Laboratory, Neuilly sur-Seine, France) (group N, n=18) was added to the lidocaine. All solutions were prepared by an anaesthesiologist who took no other part in the study. The anaesthesiologist who performed the axillary plexus blocks was unaware of the mixture that he injected.
Mean arterial pressure (MAP) and heart rate (HR) were measured using a non-invasive automated oscillometric device before surgery, every 5min during surgery and every 30min until discharge from the day-surgery unit. Duration of analgesia was defined as the time from axillary injection to the first request for supplementary analgesia. On request for supplementary analgesia, patients were given paracetamol (acetoaminophen) 500 mg plus codeine 30 mg orally. Visual analogue pain score (VAS) was assessed on a 100-mm scale every hour within the 6-h period after the axillary brachial plexus block. Postoperative follow-up was performed by a blinded observer and a telephone interview was performed on the first and second days after the axillary brachial plexus block. Subsequent 48-h consumption of analgesics and occurrence of side-effects were recorded.
In a previous study , 500 μg of intra-articularly administered neostigmine induced a prolongation of postoperative analgesia. Based on the data of this study, a power analysis indicated that a sample size of 17 patients in each group was sufficient to have an 80% chance (type II error = 0.2) of detecting a similar difference of duration of postoperative analgesia at the 95% (type I error = 0.05) significance level. Patient data were analysed using the unpaired Student's t-test. Time to the first request for supplementary analgesics was compared using ANOVA followed by the Neuman-Keuls test if differences between groups were found. Non-parametric data were analysed using two-way ANOVA for repeated measures and the Mann-Whitney U-test. Nominal data were analysed using the Χ2-test. Results are expressed as mean (SD). Statistical significance was considered at P< 0.05. All analyses were performed using the statistical package Statview (Statview for Windows and Macintosh, version 5.0, Abacus Concepts, Berkeley, CA, USA).
There were 18 patients in group N and 16 in group S. Among the patients, none was excluded from the study because of technical failure. There were no statistical differences between groups for demographic data or duration of surgery (Table 1). The duration of analgesia did not significantly differ between groups: 812.5 (456.9) min for group S vs. 746.7 (474.1) min for group N (P>0.05). None of the patients were kept more than 6h in the day-surgery unit. An interview was performed with all on the first and second days after the axillary brachial plexus block. The need for supplementary analgesia did not significantly differ between groups: 4.4 (1.5) extra doses for group S vs. 3.8 (2.2) extra doses for group N (P>0.05). As shown in Fig. 1, VAS did not significantly differ between groups, but they were only recorded within the 6h after the axillary brachial plexus block in these outpatients. Whether they would have significantly differed afterwards remains unknown.
Evaluation of adverse effects included assessment of nausea, vomiting, diarrhoea and bradycardia (defined as a 20% decrease in heart rate compared with preoperative value). No patient in group S experienced side-effects. Two patients in group N experienced nausea lasting 2-h after the axillary brachial plexus block (NS). The telephone interview conducted 24 h and 48 h after discharge from the day-surgery unit did not reveal any recurrent side-effects. Neither puncture of the axillary artery occurred nor was parasthesia elicited during axillary brachial plexus block in either group.
Our results demonstrated that neostigmine 500 μg added to 1.5% lidocaine 450mg and epinephrine 5μg mL−1 for axillary brachial plexus block did not increase the duration of postoperative analgesia in patients undergoing elective carpal tunnel release.
Preclinical studies have shown that acetylcholine receptors exist at the peripheral nerve level  and that acetylcholine induces analgesia, at least peripherally, by increasing cyclic GMP via generation of nitric oxide [5, 6, 7]. Recently, a study showed a peripheral analgesic effect of intra-articularly administered neostigmine in a model of an inflamed knee joint in rats . Finally, in a recent clinical study in patients undergoing arthroscopic meniscus repair, Yang and his colleagues showed that intra-articular administration of neostigmine 500 μg produced moderate but significant analgesia after surgery . The acetylcholine receptors are located at the level of the spinal ganglion and at the distal ending of the peripheral sensory neurone. Both Duarte  and Yang  and their colleagues administered neostigmine at the peripheral nerve ending. Therefore, the analgesic effect may have been the consequence of a pure local effect of neostigmine. Data have shown that peripheral nerves contain elements needed for muscarinic receptormediated signal transduction . Therefore, we assumed that neostigmine may act at the level of a plexus nerve trunk. Despite these preclinical and clinical data, we failed to demonstrate any prolongation of postoperative analgesia when neostigmine 500 μg was added to local anaesthetics for axillary brachial plexus block.
Hassan and his colleagues  found that peripheral inflammatory conditions enhanced the analgesic action of locally administered opioids and it has been shown that the breakdown of the perineurium by inflammation increases the accessibility of peripheral opioid receptors . Based on studies showing similarities of opioid and cholinergic systems [9, 10, 11, 12, 13, 14], it might be a reasonable assumption that the analgesic effect of neostigmine could be explained by a local effect due to inflammatory processing. Neostigmine is a highly ionized compound (pKa 12.0); therefore, the considerable lipid coverings of the nerves in the brachial plexus might be an obstacle to neostigmine to access to peripheral muscarinic receptors. Thus, in the absence of perineurium disruption due to inflammation the access to peripheral muscarinic receptors might be difficult. Therefore, the lack of analgesic action of neostigmine when administered at the level of a nerve plexus distant from the surgical site might be due to the absence of perineurium disruption. The onset of the block was not assessed. It remains unknown whether taking the onset of the block-as the reference point for the duration of analgesia - instead of the time of injection would have made any difference.
The use of spinal neostigmine is limited by sideeffects caused by its cephalad spread. Neostigmine does not cross the blood-brain barrier. Therefore, analgesia achieved by peripheral delivery of neostigmine could be attractive because it might not be limited by side-effects. In our study, neostigmine did not induce significant side-effects but it is conceivable that the limited number of patients and the low dose of neostigmine used rule out a specific conclusion. Nausea experienced by two patients in the neostigmine group could be explained by a systemic absorption of neostigmine. Many of the patients were at home when they took their first postoperative dose of analgesic drugs. Taking extra doses for analgesia in the home environment may be influenced by nonspecific factors.
In summary, we failed to demonstrate any prolongation of postoperative analgesia when neostigmine 500 μg was added to lidocaine for axillary brachial plexus block. Therefore, neostigmine does not seem to be of any clinical value for peripheral nerve blockade.
1 Yaksh TL, Grafe MR, Malkmus S et al.
Studies on the safety of chronically administered intrathecal neostigmine
methylsulfate in rats and dogs. Anesthesiology
2 Lauretti GR, Reis MP, Prado WA, Klamt JG. Doseresponse study of intrathecal morphine versus intrathecal neostigmine
, their combination, or placebo for postoperative analgesia
in patients undergoing anterior or posterior vaginoplasty. Anesth Analg
3 Hwang JH, Hwang KS, Leem JK, Han SM, Lee DM. The antiallodynic effects of intrathecal cholinesterase inhibitors in a rat model of neuropathic pain. Anesthesiology
4 Day NS, Berti-Mattera LN, Eichberg J. Muscarinic cholinergic receptor-mediated phosphoinositide metabolism in peripheral nerve. J Neurochem
5 Duarte IDG, Lorenzetti B, Ferreira S. Peripheral analgesia
and activation of nitric oxide-cyclic GMP pathway. Eur J Pharmacol
6 Ferreira SH, Nakamura M. Prostaglandin hyperalgesia. A cAMP/Ca dependent process. Prostaglandins
7 Iwamoto ET, Marion L. Pharmacologic evidence that spinal muscarinic analgesia
is mediated by an L1
arginine/nitric oxide/cyclic GMP cascade in rats. J Pharmacol Exp Ther
8 Yang LC, Chen LM, Wang CJ, Buerkle H. Postoperative analgesia
by intra-articular neostigmine
in patients undergoing knee arthroscopy. Anesthesiology
9 Buerkle H, Boschin M, Marcus MAE, Brodner G, Wüsten R, Van Aken H. Central and peripheral analgesia
mediated by the acetylcholinesterase-inhibitor neostigmine
in the rat inflamed knee joint model. Anesth Analg
10 Hassan AHS, Ableitner A, Stein C, Hertz A. Inflammation of the rat paw enhances axonal transport of opioid receptors in the sciatic nerve and increases their density in the inflamed tissue. Neuroscience
11 Antonijevic I, Mousa SA, Shafer M, Stein C. Perineural defect and peripheral opioid analgesia
in inflammation. J Neurosci
12 Gillberg PG, Aquilonius SM. Cholinergic, opioid and glycine receptor binding sites localized in human cord by in vitro autoradiography. Acta Neurol Scand
13 Gillberg PG, Wiksten B. Effects of spinal lesions and rhizotomies on cholinergic and opiate receptor binding sites in rat spinal cord. Acta Physiol Scand
14 Scatton B, Dubois A, Javoy-Agid F, Camus A. Autoradiographic localization of muscarinic cholinergic receptors at various segmental levels of human spinal cord. Neurosci Lett